EFFECT OF CONTROLLABLE VARIABLES UPON THE 
PENETRATION TEST FOR asphalts AND asphalt 
CEMENTS 
By Provost Hubbard, Chemical Engineer, and F. P. Pritchard, Assistant Chemist, 
Office of Public Roads and Rural Engineering 
INTRODUCTION 
No one test for asphalts and asphalt cements is probably better known 
or more generally used than the penetration test. Many instruments 
have been devised for determining the consistency of these materials, but 
none have been generally adopted that do not substantially conform to 
the fundamental principles of the apparatus known as the Dow penetra¬ 
tion machine. 1 This machine and others designed to give practically 
equivalent results are too well known to require description in this 
paper. In general, however, it may be said that by their use the con¬ 
sistency of asphalts or asphalt cements is expressed as the depth in 
hundredths of a centimeter that a standard needle will penetrate them 
vertically without external friction while the material is maintained at 
a stated temperature and the needle is operated under a stated load for 
a stated length of time. In the Dow penetration machine external fric¬ 
tion is practically eliminated. In other satisfactory types it is reduced 
to an almost negligible minimum, but when operating with those in 
which the needle holder slides through a guiding sleeve it is most impor¬ 
tant that both the plunger and sleeve be absolutely clean and dry, as a 
small amount of moisture, oil, or dirt will produce considerable friction 
and thus retard the penetration of the needle into the sample being 
tested. Certain standards of temperature, load, and time have been 
generally adopted, and the most widely used combination is 25 0 C., 
100 gm., 5 seconds. 
Granting that the apparatus is mechanically satisfactory and that a 
definite standard needle is used, the test appears to be comparatively 
simple. It has frequently been found, however, that different labora¬ 
tories, working upon samples of the same material under supposedly 
identical conditions of temperature, load, and time, obtain appreciably 
different results. The object of this investigation has therefore been to 
determine what effect apparently slight differences in these conditions 
will produce in the results of tests and also to study the importance of 
other controllable variables. 
1 Dow, A. W. The testing of bitumens for paving purposes. In Proc. Amer. Soc. Testing Materials, 
6th Ann. Meeting 1903, v. 3, p. 349-368, fig. 1-6. Discussion, p. 369-373- 1903• 
(8°s) 
Journal of Agricultural Research, 
Dept, of Agriculture, Washington, D. C. 
cb 
Vol. V, No. 17 
Jan. 24,1916 
D —2 
